US Crude Oil Production: Monthly EIA Series Since 1920

US crude oil production by month and by year since 1920, from the EIA, with the record 13.66 million barrels a day of 2025. CSV download, updated monthly.

US crude oil production is the volume of crude oil, lease condensate included, extracted from wells in the United States, measured in barrels per day. The US Energy Information Administration publishes it monthly from January 1920 in its Petroleum Supply Monthly, as series MCRFPUS2, and this dataset carries the full monthly record. Output averaged 13.66 million barrels a day in 2025, the highest annual level in a series the EIA starts in 1859.

TL;DR

US crude output took 48 years to beat its 1970 peak, then reached 13.66 million barrels a day in 2025, according to the EIA.

  • The annual average peaked at 9,637 thousand barrels a day in 1970, fell to 5,000 thousand in 2008 and passed the 1970 level again in 2018; 2025 was 2.73 times 2008.
  • The monthly record is 13,995 thousand barrels a day, set in October 2025; the latest monthly reading is 13,948 thousand for July 2026.
  • Between September 2008, a hurricane month, and April 2015 output rose 142.9%, the largest gain over any 79-month window since 1920.

Dataset: US Crude Oil Production, monthly EIA series (1920–2026) · Latest observation 2026-07-01

Latest Value
13,948.00
thousand barrels per day · Jul 1, 2026
Historical Percentile
99.8th
Historically high
Historical Average
6,575.82
thousand barrels per day · 1,279 observations
Historical Range
HIGH Oct 1, 2025
13,995.00
LOW Jan 1, 1920
1,097.00
thousand barrels per day

Data provided as is, with no warranty of accuracy or timeliness. Check any figure against the primary source cited on this page before using it. Terms: legal notice.

One email a month. The month's new study and the macro barometer, with links to the data behind them. Privacy policy


Loading data…

Source: US Energy Information Administration (EIA), Petroleum Supply Monthly · series MCRFPUS2


Macro Takeaway

Field production is the supply side of the US crude balance: with imports, it feeds refinery runs, exports and the change in inventories, which the EIA tracks each week in weekly US commercial crude stocks. The series also explains the emergency stockpile. The Strategic Petroleum Reserve since 1977 was filled mostly between 1977 and 2009, while domestic output peaked again at 8,971 thousand barrels a day in 1985 and then fell to its 2008 low: a stock built for a country that imported more every year. Output also answers prices with a delay, because new wells take months to drill and complete, so a month reads best against the price path of the year before.

The pandemic year shows that lag in reverse. Output fell 25.2% between November 2019 and May 2020, the steepest six-month drop on record, as operators shut in wells when demand collapsed. The recovery took longer. The annual average of 11,324 thousand barrels a day in 2021 was still below 2020, and the 2019 record of 12,335 thousand was only beaten in 2023. By October 2025 monthly output reached a record 13,995 thousand barrels a day.


Dataset Overview

IndicatorUS field production of crude oil, including lease condensate (1920–2026)
GeographyUnited States (50 states and federal offshore)
FrequencyMonthly
PeriodJanuary 1920 to the latest EIA release
Variablesdate, crude_oil_production
UnitThousand barrels per day
FormatCSV, Excel (XLSX), JSON
SourcesUS Energy Information Administration (EIA), Petroleum Supply Monthly, series MCRFPUS2
Last updated2026-07-01

Dataset Variables

The CSV and Excel files contain the following columns.

ColumnTypeDescription
dateDate (YYYY-MM-DD)First day of the month the observation covers
crude_oil_productionIntegerUS field production of crude oil, including lease condensate, thousand barrels per day (EIA MCRFPUS2)

Column names match the CSV headers exactly. Annual averages quoted on this page are the EIA annual series, which weight months by their number of days.


Change Between Two Months, Ranked Since 1920

A production change only means something against the changes the series has already seen. Pick a start month and an end month. The tool returns the change in thousand barrels a day and in percent, then ranks it against every window of the same length since January 1920. A 24-month gain is compared with all other 24-month gains, a decline with all other declines.

The default window, September 2008 to April 2015, covers the first shale surge: 5,678 thousand barrels a day added in 79 months, the largest of the 1,200 windows of that length in the record. Its starting point is depressed. In September 2008, Hurricanes Gustav and Ike shut down almost all Gulf of Mexico output, according to the EIA: federal offshore production fell to 242 thousand barrels a day, from 1,272 thousand in August 2008. Measured from August 2008, the rise to April 2015 is 4,649 thousand barrels a day, or 92.9%. The pandemic window, November 2019 to May 2020, is the steepest fall of all 1,273 six-month windows.


Download the Complete Dataset

The full US crude oil production dataset is available in CSV and Excel formats.

Data provided as is, with no warranty of accuracy or timeliness. Check any figure against the primary source cited on this page before using it. Terms: legal notice.

One email a month. The month's new study and the macro barometer, with links to the data behind them. Privacy policy


Direct CSV Access: Eco3min Structured Dataset

https://eco3min.fr/dataset/eia/us-crude-oil-production.csv

This URL returns the complete dataset in CSV format. It can be used directly in pandas, R, curl, or any data tool. A JSON version is served at the same path with the .json extension. The underlying data is EIA series MCRFPUS2; this file gives it a stable header and month-start dates.


Using the Dataset in Python

import pandas as pd

url = "https://eco3min.fr/dataset/eia/us-crude-oil-production.csv"
df = pd.read_csv(url, parse_dates=["date"])
annual = df.groupby(df["date"].dt.year)["crude_oil_production"].mean()

print(df.tail())
print(annual.idxmax(), round(annual.max()))

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/dataset/eia/us-crude-oil-production.csv"
df <- read_csv(url)
df$year <- as.integer(format(df$date, "%Y"))

tail(df)
aggregate(crude_oil_production ~ year, data = df, FUN = mean)

Both examples load the dataset directly from the URL, with no download or API key required. A simple mean of months differs slightly from the EIA annual average, which weights each month by its number of days: 11,355 against 11,349 thousand barrels a day for 2020.


Methodology

The EIA publishes monthly field production in the Petroleum Supply Monthly at the end of each month, two months after the month covered: the release of 30 September 2026 carried July 2026. Lease condensate, a light liquid recovered with natural gas at the wellhead, is included. An Eco3min pipeline pulls the series from the EIA Open Data API v2 (route petroleum/crd/crpdn) after each release and rewrites the CSV, Excel and JSON files from the full history.

Recent months can be revised in later releases, which is why the file is rebuilt in full each time. The EIA also publishes a weekly estimate in its Weekly Petroleum Status Report (series WCRFPUS2, from 1983), which reached 13,979 thousand barrels a day in the week to 2 October 2026, the highest in that series; this dataset keeps the monthly series only.


Historical Regimes

  • 1920–1970: the climb to the 1970 peak. Monthly output rose from 1,097 thousand barrels a day in January 1920 to an annual average of 9,637 thousand in 1970, the decade averages climbing from 1,949 thousand in the 1920s to 7,995 thousand in the 1960s. The single highest month of that era, November 1970, reached 10,044 thousand barrels a day.
  • 1971–1985: Alaska offsets the Lower 48. Production in the lower 48 states fell after 1970, and the national total dropped to 8,132 thousand barrels a day in 1976. The Prudhoe Bay field then came on stream: Alaska went from 173 thousand barrels a day in 1976 to 464 thousand in 1977 and 1,229 thousand in 1978, lifting the US total back to 8,971 thousand by 1985, still below the 1970 level.
  • 1986–2008: the long decline. With prices lower after 1986 and Alaska past its 1988 peak of 2,017 thousand barrels a day, output fell almost every year, to 5,822 thousand in 2000 and 5,000 thousand in 2008, the lowest annual level since 1946 and 48.1% below 1970.
  • 2009–2015: the shale surge. Horizontal drilling and hydraulic fracturing opened tight formations, notably in Texas, North Dakota and New Mexico. Between September 2008 and April 2015 monthly output rose by 5,678 thousand barrels a day, or 142.9%: the largest increase over any 79-month window in the record. Texas alone went from 1,109 thousand barrels a day in 2008 to 5,761 thousand in 2025.
  • 2016–2019: price slump, then a new record. The oil price collapse of late 2014 cut the annual average from 9,433 thousand barrels a day in 2015 to 8,859 thousand in 2016. Output recovered quickly: in November 2017 a month beat the 1970 monthly high for the first time, and the annual average of 10,972 thousand in 2018 passed the 1970 level, 48 years later. 2019 reached 12,335 thousand.
  • 2020–2026: pandemic drop and successive records. Output fell from 13,019 thousand barrels a day in November 2019 to 9,734 thousand in May 2020, a 25.2% drop and the steepest six-month decline in the series. Annual output then set records in 2023, 2024 and 2025, at 12,972, 13,267 and 13,662 thousand barrels a day.

The where the US ranks among oil producers page sets these regimes against the rest of the world. In the EIA’s international data the United States was the largest producer in 1973, ahead of the Soviet Union, and it has been again every year since 2018.


Related Macroeconomic Datasets

Related Research

Deep Analytical Framework

The structural reading of oil, gas and critical minerals, from supply constraints to price regimes, is set out in physical commodity markets.


Macroeconomic Dataset Hub

This dataset is part of the Eco3min macro-financial data repository. Explore all available datasets, including energy benchmarks, interest rates, inflation and credit series.

Explore the Eco3min Dataset Hub


Sources

  • US Energy Information Administration (EIA), Petroleum Supply Monthly, US field production of crude oil, thousand barrels per day (MCRFPUS2), via the EIA Open Data API v2, release of 30 September 2026
  • US Energy Information Administration (EIA), field production by state and area (Alaska MCRFPAK2, Texas MCRFPTX2), and weekly production estimates (WCRFPUS2), Weekly Petroleum Status Report
  • Eco3min Research, structured dataset compilation, window ranking and regime dating

EIA data are US government publications in the public domain; the EIA asks users to cite the source and the publication date.


Dataset Reference


Last updated — 14 October 2026

Disclaimer – Financial and Legal Information: The analyses, commentary, and content published on eco3min.fr are provided for informational and educational purposes only. They do not constitute investment advice or a solicitation to buy or sell financial instruments. Past performance is not indicative of future results. All investment decisions involve risk and are the sole responsibility of the reader. Data, charts and calculated series are provided as is and may contain errors or lag behind their source: check any figure against the primary source cited before relying on it. Content describing laws, regulations or contracts reflects the law as of its publication date and does not constitute legal or tax advice; applying it to a specific situation requires a qualified legal or tax professional.

Data on this page: the series are collected automatically from the sources cited, then formatted or calculated by Eco3min (units, frequency, derived calculations). They may differ from the official release because of producer revisions, update lags, rounding or processing errors. They are provided as is, with no warranty of accuracy, completeness or timeliness. Before any professional, contractual or regulatory use, and before any financial decision, check the values against the primary source. To the extent permitted by law, the publisher accepts no liability for the consequences of using these data without that check.